CN112783172B

CN112783172B - AGV and machine integrated scheduling method based on discrete whale optimization algorithm

Info

Publication number: CN112783172B
Application number: CN202011632120.5A
Authority: CN
Inventors: 邹裕吉; 宋豫川; 王馨坤
Original assignee: Chongqing University
Current assignee: Chongqing University
Priority date: 2020-12-31
Filing date: 2020-12-31
Publication date: 2023-03-31
Anticipated expiration: 2040-12-31
Also published as: CN112783172A

Abstract

The invention discloses an AGV and machine integrated scheduling method based on a discrete whale optimization algorithm, which utilizes the advantages of simple structure, few parameters, strong searching capability, easiness in implementation and the like of the whale optimization algorithm, greedy decoding is carried out on each whale individual in a whale population by combining the running path of an AGV trolley and the corresponding transportation running time in the iterative updating process of the whale optimization algorithm, so that the integrated scheduling problem of the AGV and the machine can be effectively solved, the scheduling scheme is optimized by optimizing the whale individual through iterative circulation, the scheduling efficiency of the integrated operation process of the AGV and the machine can be improved, the overall searching capability and the local optimal jumping capability of the algorithm can be further enhanced by means of operations such as whale population iterative updating treatment introducing levy flight strategies, local searching strategies adopting a plurality of optimal whale individuals and the like, the overall searching capability and the local convergence accuracy of the algorithm can be enhanced, and good solving and operation efficiency can be kept.

Description

AGV and machine integrated scheduling method based on discrete whale optimization algorithm

Technical Field

The invention relates to the technical field of workshop scheduling, in particular to an AGV and machine integrated scheduling method based on a discrete whale optimization algorithm.

Background

With the development of information technologies such as cloud computing, internet, big data and the like, large-scale customized and multi-variety small-batch production modes gradually become mainstream. Under the production mode mainly based on individuation, the products are various in types, the process flows are different, and a reasonable workshop scheduling scheme is crucial to improving the production efficiency. In the traditional flexible job shop scheduling problem research, the workpiece transfer time is not generally considered, so that the scheduling result is not a theoretical optimal scheduling result, and the method has a defect in guiding actual production. An Automated Guided Vehicles (AGV) is an advanced logistics equipment with high flexibility, high reliability and high efficiency, and the AGV is used for transferring workpieces in many production workshops. In this case, the dispatching of the machine and the dispatching of the AGV can influence each other, so the research of the integrated dispatching problem of the AGV and the machine has important theoretical value and engineering significance.

In recent years, a great deal of research is carried out by domestic and foreign scholars on the integrated scheduling of the AGVs and the machines, most of the scholars are focused on solving the problems of task assignment and task execution time sequence of the AGVs and the machines, the driving roads of the AGVs are set to be one-way, and potential path conflicts are not considered. However, when the traveling road of the AGV is one-way and two-way, the operation efficiency of the system will be improved, but the path conflict will also increase. If the path conflict is not solved, the problems of AGV collision, dead-locked path and the like can occur, the scheduling plan is disturbed, and even the production system is paralyzed.

The problem of AGV and machine integrated scheduling has been proven to be the superposition of NP difficult problems, the solution space of the problems is large and complex, the accurate algorithm is difficult to obtain the optimal solution within the acceptable time, and the heuristic algorithm is more suitable for the large-scale NP difficult problems. Therefore, how to find a solution to the problem of integrated scheduling of AGVs and machines becomes an emerging research topic in the field.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides a novel solution for an AGV and machine integrated scheduling method based on a discrete whale optimization algorithm, so as to effectively solve the problem of AGV and machine integrated scheduling and help improve the scheduling efficiency of an AGV and machine integrated operation process.

In order to solve the technical problems, the invention adopts the following technical scheme:

an AGV and machine integrated scheduling method based on a discrete whale optimization algorithm comprises the following steps:

acquiring scheduling task parameters of an integrated operation process of an AGV and a machine; carrying out whale individual coding processing of a whale optimization algorithm according to the scheduling task parameters to generate whale populations with preset population scales, wherein each whale individual corresponds to an AGV and machine integrated operation flow scheduling scheme; carrying out iterative updating on the whale population by adopting a whale optimization algorithm, carrying out greedy decoding on each whale individual in the whale population by combining the running path of the AGV trolley and the corresponding transportation running time in the iterative process, and selecting an optimal whale individual; and executing the operation flow scheduling of the AGV trolleys and the machines according to the AGV and machine integrated operation flow scheduling scheme corresponding to the optimal whale individual.

In the AGV and machine integrated scheduling method based on the discrete whale optimization algorithm, the scheduling task parameters preferably include processing parameters of workpieces, transportation operation parameters of an AGV trolley and map information of a transfer transportation area in an operation process.

In the above method for integrated scheduling of AGVs and machines based on discrete whale optimization algorithm, preferably, the specific process of the method includes the following steps:

1) Acquiring scheduling task parameters, and setting initialization parameters of a whale optimization algorithm; the initialization parameters comprise a population size NIND and a maximum iteration number;

2) Carrying out whale individual coding processing of a whale optimization algorithm according to the scheduling task parameters to generate whale populations with the NIND whale individual scales;

3) Judging whether the iteration loop times reach the maximum iteration times, and if so, skipping to execute the step 9); otherwise, planning running paths of all AGV dollies in the integrated operation flow scheduling scheme of the AGVs and the machines corresponding to all whale individuals in the whale population, determining the transportation running time of each AGV dollies in the planned path, carrying out greedy decoding on each whale individual in the whale population by combining the running paths of the AGV dollies and the corresponding transportation running time, respectively calculating the maximum completion time of the integrated operation flow scheduling scheme of the AGVs and the machines corresponding to each whale individual in the whale population, recording the whale individual with the shortest current global maximum completion time as a global optimal whale individual, and executing the step 4);

4) Judging whether the iteration times of the currently recorded globally optimal whale individual, which are kept unchanged, reach a preset algebraic threshold value or not; if yes, executing step 5), otherwise, executing step 6);

5) Judging whale individuals to be inferior according to the maximum completion time, randomly generating new whale individuals with the population size of 50% to replace 50% of inferior whale individuals in the whale population, and executing the step 6);

6) Carrying out whale optimization algorithm iterative updating on the whale population;

7) For n in the current whale population _x Carrying out local search on individual superior whales; n is _x A preset local search quantity parameter;

8) Keeping the whale population scale, selecting a superior whale individual, entering the next iteration, and returning to the step 3);

9) And taking the globally optimal whale individual recorded at present as the selected optimal whale individual, and executing the operation flow scheduling of the AGV trolley and the machine according to the AGV and machine integrated operation flow scheduling scheme corresponding to the optimal whale individual.

In the above method for integrated scheduling of AGVs and machines based on discrete whale optimization algorithm, preferably, the whale individual includes a plurality of genes, each gene corresponds to one operation process in the AGV and machine integrated operation flow scheduling scheme, so that a whale individual is formed by a set of genes corresponding to each operation process in the AGV and machine integrated operation flow scheduling scheme;

the whale individual codes of all genes in the whale individuals comprise procedure code segments, machine code segments and AGV code segments, wherein the procedure code segments are used for indicating workpiece numbers corresponding to the operation procedures, the machine code segments are used for indicating machine numbers corresponding to the workpiece numbers indicated in the procedure code segments in the operation procedures, and the AGV code segments are used for indicating AGV numbers corresponding to the workpiece numbers indicated in the procedure code segments in the operation procedures; and in a whale individual, the repeated sequence of the genes with the same process code segment is used for indicating the processing process of the workpiece corresponding to the workpiece number in the corresponding process code segment.

In the method for integrated dispatching of the AGV and the machine based on the discrete whale optimization algorithm, preferably, each code segment of the gene in the whale individual is a continuous code with a value interval of [ -delta, delta ], and a discrete code is obtained after discrete code conversion so as to indicate a workpiece number, a machine number and an AGV number corresponding to the gene;

the discrete transcoding method for individual whale is as follows:

carrying out discrete code conversion by adopting a maximum position value rule aiming at the process code segment of each gene in the whale individual;

aiming at the machine code segment and the AGV code segment of each gene in the whale individual, discrete code conversion is carried out respectively in the following modes:

m _equ (i) Representing continuously encoded machine code segment values, z _equ (i) Indicates the number of processing machines selectable for the corresponding process, u _equ (i) Representing the resulting discretely encoded machine code segment values; m is _AGV (i) Representing successively encoded AGV segment values, z _AGV (i) Indicates the number of AGV carts selectable for the corresponding process, u _AGV (i) Representing the obtained discrete coded AGV segment value; delta is the value range [ -delta, delta ] of continuous coding]The upper limit end value of (1); run [ 2 ]]Rounding the rounding operator.

In the above AGV and machine integrated scheduling method based on discrete whale optimization algorithm, preferably, in step 3), the constraint condition for greedy decoding of each individual whale in the whale population includes:

constraint (1): any one procedure can be processed only by one machine;

constraint (2): at most one AGV is responsible for transportation in any process;

constraint condition (3): the AGV no-load starting time is not earlier than the last transport task ending time and the workpiece starting time of the AGV;

constraint (4): the AGV no-load ending time is the sum of the no-load starting time and the no-load running time;

constraint (5): the AGV load starting time is not earlier than the AGV no-load ending time and the workpiece completion time;

constraint (6): the AGV load ending time is the sum of the load starting time and the load running time;

constraint (7): the starting time of the process is not earlier than the finishing time of the load and the finishing time of the preorder process of the machine;

constraint (8): the finishing time of the working procedure is the sum of the start time and the processing time;

constraint (9): the completion time of the workpiece is the time for the AGV to convey the workpiece to the warehouse;

constraint (10): when an AGV enters a certain road section at a certain moment, the AGV is not allowed to drive into the road section from an outlet of the AGV before driving out of the road section;

constraint (11): any position node can only hold the next AGV at the same time.

In the above AGV and machine integrated scheduling method based on the discrete whale optimization algorithm, preferably, in the step 3), the whale individual with the shortest current global maximum completion time is the whale individual with the shortest maximum completion time in each iteration from the current iteration time.

In the above AGV and machine integrated scheduling method based on discrete whale optimization algorithm, preferably, in step 7), the neighborhood structure adopted by the local search includes:

neighborhood structure 1: randomly selecting two process code sections of each gene in the whale individual, wherein the two process code sections are required to correspond to the processes of different workpieces, and interchanging the positions of the two process code sections;

neighborhood structure 2: randomly selecting two process code sections of each gene in the whale individual, and inserting the next process code section in front of the previous process code section;

neighborhood structure 3: randomly selecting one of the process code segments of each gene in the whale individual, wherein the number of machinable machines in the process corresponding to the process code segment is more than 1, and randomly selecting one machining machine for replacing the process corresponding to the process code segment from the machinable machines in the process corresponding to the process code segment.

In the above method for integrated scheduling of AGVs and machines based on discrete whale optimization algorithm, preferably, in step 7), the specific manner of locally searching each whale individual is as follows:

the variable neighborhood search based on the process and the machine is firstly carried out according to the following steps:

step a1: firstly, X is _a Updating the whale individuals subjected to local search currently; let the current iteration number n _a =1, let maximum number of iterations n _a,max =5, let p _a 1, let p _a,max ＝3；

Step a2: judging whether a circulation termination condition n is reached _a ≥n _a,max (ii) a If it reaches, output X _a A corresponding individual whale; otherwise, turning to the step a3;

step a3: at X _a Randomly selecting a neighborhood structure on the basis of the corresponding whale individual to obtain a disturbed individual X' _a ；

Step a4: in perturbing individual X' _a The variable neighborhood searching is carried out on the basis, and the specific steps are as follows:

a4.1 ) whether or not the termination condition p is reached is judged _a ≥p _a,max If so, output the current descrambled individual X' _a Turning to the step a5; otherwise, turning to the step a 4.2);

a4.2 X 'in a perturbed individual' _a Based on the selection number and the current p _a Get the new individual X ″' with the neighborhood structure corresponding to the value _a If f (X ″) _a )＜f(X′ _a ) Then X 'is updated' _a ←X″ _a ，p _a ← 1; if f (X ″) _a )＝f(X′ _a ) Then X 'is updated with a probability of 0.5' _a ←X″ _a ，p _a ← 1; otherwiseX′ _a Not updated, p _a ←p _a +1; then return to a 4.1);

step a5: updating X _a ←X′ _a ，n _a ←n _a +1, turning to the step a2;

wherein, f (X' _a )、f(X″ _a ) Respectively represent individual X' _a 、X″ _a Respectively corresponding maximum completion time of machining;

then, neighborhood search based on the constraint AGV is carried out according to the following steps:

step b1: mixing X _b Updating X output for the variable neighborhood search based on process and machine _a Corresponding individual whale; order the AGV task number n of the current operation _b Is 1, let n _b,max The total task number of the AGV currently operating;

step b2: judging whether a termination condition n is satisfied _b ≥n _b,max If satisfied, output X _b The corresponding whale individual is used as a preferred object of the whale population in the step 8); otherwise, turning to the step b3;

and b3: distributing the AGV task operated currently to the AGV trolley with the minimum current transportation running time to obtain a new individual X' _b (ii) a If f (X' _b )＜f(X _b ) Update X _b ←X′ _b (ii) a If f (X' _b )＝f(X _b ) Then X is updated with a probability of 0.5 _b ←X′ _b (ii) a Otherwise, X is not updated _b (ii) a Then n is updated _b ←n _b +1, go to step b2;

wherein f (X' _b )、f(X″ _b ") respectively denote individual X' _b 、X″ _b The respective AGVs transport the maximum completion time.

In the above AGV and machine integrated scheduling method based on discrete whale optimization algorithm, as preferable, in step 6), the specific manner of performing whale optimization algorithm iterative update on whale population is as follows:

updating a calculated coefficient vector

The value of (c):

convergence factor a =2- (2 t)/t _max ；

Wherein t is the current iteration number, t _max Is the maximum number of iterations, r ₁ 、r ₂ Are all [0,1]A random number between values; in [0,1]Randomly taking values to generate a probability parameter p;

when coefficient vector

Is greater than or equal to>

And p is<At 0.5, whale individual position updating is carried out according to the following mode:

the first method is as follows:

when coefficient vector

Is greater than or equal to>

And when p is more than or equal to 0.5, updating the positions of the whale individuals according to the following second mode:

the second method comprises the following steps:

when coefficient vector

Is greater than or equal to>

And then, updating the position of the whale individual according to the following three ways:

the third method comprises the following steps:

wherein the content of the first and second substances,

indicating the updated individual position of the whale; d represents an update step size; />

Indicating the individual position of the present whale>

Location vector representing a randomly selected individual whale, <' > based on the location of the whale>

Representing a target prey location; b is a defined logarithmic spiral shape constant; l is [ -1,1]A random number taken between; rand is subject to a range of [0,1]A uniform distribution function of; when/is>

Then, [ rand-1/2 ]]Taking-1; when/is>

Then, [ rand-1/2 ]]Taking 0; when +>

Then, [ rand-1/2 ]]Taking 1; />

Represents the inner product operation of the matrix, levy represents the Levy flight factor, and:

Levy(s)～|s| ^-1-β ,0＜β≤2；

u, v and beta are Levy flight parameters; u obeys normal distribution

v obey a normal distribution>

Γ is the standard gamma function.

Compared with the prior art, the invention has the beneficial effects that:

1. according to the AGV and machine integrated scheduling method based on the discrete whale optimization algorithm, the advantages of simple structure, few parameters, strong searching capability, easiness in implementation and the like of the whale optimization algorithm are utilized, greedy decoding is performed on each whale individual in a whale population by combining the running path of an AGV trolley and the corresponding transportation running time in the iterative updating process of the whale optimization algorithm, so that the AGV and machine integrated scheduling problem is effectively solved, the whale individual is optimized through iterative circulation to achieve optimization of a scheduling scheme, and the scheduling efficiency of the AGV and machine integrated operation process can be improved.

2. According to the AGV and machine integrated scheduling method, by means of whale population iterative updating processing of a levy flight strategy and cycle processing operation based on overall optimal whale individual maintenance algebra, overall searching capacity and local optimal jumping capacity of an algorithm can be enhanced, and scheduling efficiency of an AGV and machine integrated operation process is improved in an optimized mode.

3. According to the AGV and machine integrated scheduling method, the convergence accuracy of the algorithm can be enhanced by means of a local search strategy of a plurality of better whale individuals.

4. The AGV and machine integrated scheduling method provided by the invention can effectively solve the AGV and machine integrated scheduling problem and improve the scheduling efficiency, and meanwhile, good solving operation efficiency is still maintained without sacrificing operation complexity.

Drawings

FIG. 1 is a flow chart of an embodiment of an AGV and machine integrated scheduling method based on a discrete whale optimization algorithm.

FIG. 2 is a diagram of an example of sequential encoding of individual whales in the method of the present invention.

FIG. 3 is a schematic diagram of discrete code conversion of individual handling steps of whale in the method of the present invention.

FIG. 4 is a schematic diagram of discrete code conversion of machine code segments of individual whales in the method of the present invention.

FIG. 5 is a schematic diagram of discrete code conversion of AGV segments of whale individuals in the method of the present invention.

FIG. 6 is a Gantt chart of the optimal scheduling result in the flexible arithmetic experiment of the present invention.

FIG. 7 shows time windows of each segment in a flexible example experiment according to the present invention.

FIG. 8 is a graph of the convergence of the algorithm in the flexible example experiment of the present invention.

Detailed Description

The invention is further illustrated by the following figures and examples.

The invention provides an AGV and machine integrated scheduling method based on a discrete whale optimization algorithm, which comprises the following steps:

acquiring scheduling task parameters of an integrated operation process of an AGV and a machine;

carrying out whale individual coding treatment on the whale optimization algorithm according to the scheduling task parameters to generate a whale population with a preset population scale, wherein each whale individual corresponds to an AGV and machine integrated operation process scheduling scheme;

performing iterative updating on the whale population by adopting a whale optimization algorithm, performing greedy decoding on each whale individual in the whale population by combining the running path of the AGV trolley and the corresponding transportation running time in the iterative process, and selecting an optimal whale individual;

and executing the operation flow scheduling of the AGV trolleys and the machines according to the AGV and machine integrated operation flow scheduling scheme corresponding to the optimal whale individual.

In the AGV and machine integrated scheduling method based on the discrete whale optimization algorithm, advantages of simple structure, few parameters, strong searching capability, easiness in implementation and the like of the whale optimization algorithm are utilized, greedy decoding is performed on each whale individual in a whale population by combining the running path of an AGV trolley and corresponding transportation running time in the whale optimization algorithm iterative updating process, so that the AGV and machine integrated scheduling problem can be effectively solved, optimization of a scheduling scheme is realized by optimizing the whale individual through iterative circulation, and scheduling efficiency of an AGV and machine integrated operation process can be improved.

Specifically, the specific flow of the AGV and machine integrated scheduling method based on the discrete whale optimization algorithm of the present invention is shown in fig. 1, and includes the following steps:

1) Acquiring scheduling task parameters, and setting initialization parameters of a whale optimization algorithm; the initialization parameters comprise a population size NIND and a maximum iteration number it _max 。

The scheduling task parameters may include processing parameters of the workpiece, transport operation parameters of the AGV, and map information of the transfer transport area in the work flow. The processing parameters of the workpieces can include the total number of machines and the total number of workpieces involved in the workpiece processing in the integrated operation flow of the AGV and the machines, and the total number of processing procedures, the total number of batches, the total number of processing batches and the like of each workpiece, and are used as basic parameters for solving the workpiece processing procedures and the machine scheduling schemes; the transport operation parameters of the AGV can comprise the total number of the AGV in the transport operation of the AGV and the machine integration operation process, the running speed of the AGV and other information, the map information of the transfer transportation area in the operation process can comprise the positions of the entrances and exits of the transfer transportation area, the placement positions of the machines in the transfer transportation area, the positions of all nodes for the AGV to run in the transfer transportation area and other information, and the transport operation parameters of the AGV and the map information of the transfer transportation area are used as basic parameters for solving the transport operation scheduling scheme of the AGV.

2) And carrying out whale individual coding treatment of a whale optimization algorithm according to the scheduling task parameters to generate whale populations with the NIND whale individual scales.

3) Judging whether the iteration cycle number it reaches the maximum iteration number it _max If yes, skipping to execute the step 9); otherwise, planning running paths of all AGV dollies in the integrated operation flow scheduling scheme of the AGVs and the machines corresponding to all whale individuals in the whale population, determining the transportation running time of each AGV dollies in the planned path, carrying out greedy decoding on each whale individual in the whale population by combining the running paths of the AGV dollies and the corresponding transportation running time, respectively calculating the maximum completion time of the integrated operation flow scheduling scheme of the AGVs and the machines corresponding to each whale individual in the whale population, recording the whale individual with the shortest current global maximum completion time as the globally optimal whale individual, and executing the step 4). The whale individual with the shortest current global maximum completion time refers to the whale individual with the shortest maximum completion time in each iteration from the current iteration times.

4) Judging whether the iteration times of the globally optimal whale individual which is recorded currently and remains unchanged reaches a preset algebraic threshold value or not; if yes, step 5) is executed, otherwise, step 6) is executed.

5) Judging that whale individuals are worse according to longer maximum completion time, randomly generating new whale individuals with the population size of 50% to replace the whale individuals with the population size of 50% worse, and executing the step 6).

6) And carrying out whale optimization algorithm iterative updating on the whale population.

7) For n in the current whale population _x Carrying out local search on individual superior whales; n is _x A preset local search quantity parameter.

8) Keeping the whale population scale, selecting a superior whale individual to enter next iteration, and returning to the step 3).

The following describes each link of the AGV and machine integrated scheduling method of the present invention.

Aiming at whale individual codes, in the method, whale individuals comprise a plurality of genes, each gene corresponds to one operation process in the AGV and machine integrated operation flow scheduling scheme, and therefore a whale individual is formed by a set of genes corresponding to the operation processes in the AGV and machine integrated operation flow scheduling scheme. During coding, the whale individual codes of each gene in the whale individuals comprise a process code segment, a machine code segment and an AGV code segment, wherein the process code segment is used for indicating the workpiece number corresponding to the operation process, the machine code segment is used for indicating the machine number corresponding to the workpiece number indicated in the process code segment in the operation process, and the AGV code segment is used for indicating the AGV number corresponding to the workpiece number indicated in the process code segment in the operation process; and in a whale individual, the repeated sequence of the appearance of the genes with the same process code segment is used for indicating the processing procedure of the workpiece corresponding to the workpiece number in the corresponding process code segment, namely, the nth time of the appearance of the genes with the same process code segment in the whale individual is indicated for executing the nth processing procedure of the workpiece corresponding to the workpiece number in the corresponding process code segment.

Because the encoding of the whale optimization algorithm is a continuous encoding mode, and the solution space is a continuous value space, the initial encoding in the whale individual and the encoding in the whale individual after iterative update are continuous encoding, and the encoding value interval is [ -delta, delta ]. In the invention, considering the iterative update of whale population, if the value of delta is too small, the relative size change of the updated whale individual process sequence gene value is large, so that whale individuals are damaged; if the value of delta is too large, the updating effect is not obvious; therefore, the value of δ is generally a number between 2 and 5. Fig. 2 shows a preferred example of encoding.

However, for the problem of integrated scheduling of AGVs and machines, the scheduling parameters of the processes, machines, and AGVs are all discrete parameters, and discrete codes obtained by converting the continuous codes into discrete codes are required to indicate discrete scheduling parameters such as workpiece numbers, machine numbers, and AGVs corresponding to genes. Therefore, in the present invention, when the decoding process is required for the genetic individuals in the whale population, a discrete transcoding process is required.

In the invention, the discrete code conversion method for whale individuals is as follows:

and (3) performing discrete code conversion by adopting an LPV (Large Position Value) rule aiming at the process code segment of each gene in whale individuals. The discrete transcoding process of the LPV rule is: firstly, setting a fixed ID corresponding to each gene in whale individuals in an increasing order from 1, then sequencing and rearranging each gene in the whale individuals according to the sequence of the values of the process code segments from large to small, disorder the fixed ID sequencing of each gene, forming mapping ID sequencing, and obtaining a new gene sequencing sequence; and then, sequentially inserting each process of the workpiece 1 into the gene position of the corresponding sequence according to the sequence of the values of the mapping IDs from small to large, then inserting each process of the workpiece 2, namely … …, and so on until all the processes of the workpieces are inserted, wherein the inserted values are the workpiece number values corresponding to the processes, and therefore, the discrete coding sequence of the process code segments in the whale individual is obtained. The specific process is shown in fig. 3, in the example shown in fig. 3, the workpiece 1 has 3 processes, the workpiece 2 has 2 processes, and the workpiece 3 has 4 processes, and the mapping IDs are sorted as (8,5,4,1,9,3,2,7,6); therefore, when inserting the workpiece number value corresponding to the process, the workpiece 1 has 3 processes, and the workpiece number 1 should be inserted into three positions with mapping IDs 1, 2 and 3, respectively, at this time, the state of the discrete coding sequence of the process code segment is represented as (X, 1, X), and "X" represents the position where the workpiece number value is not inserted; then, the work 2 has 2 processes, and should be mapped to two bits with ID 4 and ID 5Inserting the workpiece number 2 respectively, wherein the state of the discrete coding sequence of the process code segment is (X, 2,1, X,1, X); finally, the work piece 3 has 4 processes, the work piece number 3 should be inserted into four positions with mapping IDs of 6, 7, 8 and 9, and the state of the discrete coding sequence of the process code segment is represented as (3,2,2,1,3,1,1,3,3), so that the discrete coding sequence of the process code segment of each gene in the whale individual is obtained. The resulting repeat sequence of gene sequence segments in individual whales, which appears from left to right, represents the processing steps in the scheduling scheme, e.g., the first 3 from left to right represents the 1 st step O of the workpiece 3 ₃₁ The second 3 from left to right represents the 2 nd step O of the work 3 ₃₂ And so on.

Aiming at the machine code segment and the AGV code segment of each gene in the whale individual, the following modes are respectively adopted for discrete code conversion:

m _equ (i) Representing continuously encoded machine code segment values, z _equ (i) Indicates the number of processing machines, u, selectable for the respective process _equ (i) Representing the resulting discretely encoded machine code segment values; m is _AGV (i) Value of AGV code segment, z, representing sequential encodings _AGV (i) Indicates the number of AGV carts selectable for the corresponding process, u _AGV (i) Representing the obtained discrete coded AGV segment value; delta is the value range of continuous coding [ -delta, delta]The upper limit end value of (2); run [ 2 ]]Rounding the rounding operator. Meanwhile, mapping from discrete space to continuous space can also be realized through inverse operation. In the AGV and machine gene string, from left to right, the machine and AGV corresponding to each process from workpiece 1 to the last workpiece are provided. See in particular fig. 4 and 5. Aiming at the problem of co-scheduling of a machine and an AGV, the coding and the conversion mode thereof cannot generate illegal solutions.

In the iterative process of the algorithm, certain dimensions of individual whales can be out of range. In the original whale optimization algorithm, after one iteration is finished, the boundary-crossing whale individual is adjusted. The handling method can cause problems because the whale individuals which cross the border are possibly selected as the basis of position updating in the whale individual iteration process, so that more whale individuals cross the border, the running time of the algorithm is increased, some whale individuals are damaged, the border-crossing dimension is set to be the maximum value or the minimum value, and when the border-crossing condition is more, the similarity among the whale individuals is larger and larger. Therefore, the invention adjusts the boundary crossing of the individual whale after the position of the individual whale is updated, and sets the boundary crossing dimension to be a value close to the boundary. The adjustment rule is shown as follows:

r ₁ represents [0,1]X (i) represents the encoded value of the ith dimension of the whale individual X.

Regarding the initialization of the whale population, when the initial whale population is generated, the quality of the population can be greatly improved by considering the working time and load balance of each machine and AGV. On the FJSP Problem (Flexible Job-shop Scheduling Problem), whale individuals with a population size of 20% are generated by using a global selection method and whale individuals with a population size of 30% are generated by using a local selection method, considering the working time and load balance on the basis of randomly generating the whale individuals with a population size of 50%. The method is also applied to the selection of the AGV, and in order to reduce the calculation amount when the AGV sequence is initialized, the conflict is not considered when the running time of the AGV is calculated.

Since there is no prior knowledge of the global optimal solution to the optimization problem, the initial population should be distributed as evenly as possible in the solution space. Therefore, the foundation of enhancing the diversity of the initial population to lay the global search of the algorithm is always one direction of algorithm optimization. At present, most researches mainly adopt a chaotic mapping strategy and an opponent learning strategy to enhance the diversity of an initial population of an intelligent optimization algorithm. Chaotic mapping and opponent learning have met with much success in improving algorithms, but since there are many chaotic mapping and opponent learning methods available to improve algorithms, it is a considerable problem to choose what method will improve the algorithms to the greatest extent.

In consideration of the above factors, the invention generates the initialized population by using the strategies of Gaussian mapping and standard opponent learning and generating the proportion of the population size of 25% respectively on the basis of randomly generating 50% whale individuals of the population size.

The gaussian mapping expression is:

the standard opponent learning expression is:

wherein the content of the first and second substances,

representing the encoded value of the ith dimension in the individual sperm of a whale after Gaussian mapping]It is shown that the rounding-off is performed,

represents the coded value of the ith dimension in the whale individual after the opposite learning, and is/are selected>

And &>

Are respectively provided withFor the coded value of the i-th dimension in randomly generated individual whales->

Upper and lower limit values of the value.

In the process of carrying out greedy decoding on each whale individual in the whale population, in order to ensure that process constraints are met among machining processes of each machine, among transportation and operation paths of each AGV trolley and between the machining processes and transportation and operation scheduling of the AGV trolleys, the following constraint conditions are preferably set:

constraint (1): any one procedure can be processed only by one machine;

constraint (3): the AGV no-load starting time is not earlier than the last transport task ending time and the workpiece starting time of the AGV;

constraint (8): the completion time of the working procedure is the sum of the start time and the processing time;

constraint (11): any position node can only hold the next AGV at the same time.

In greedy decoding, the optimal solution is necessarily in the active scheduling set when the maximum completion time is the minimum as the optimization target. Therefore, active greedy decoding is adopted, and on the premise that the arranged working procedure start time is not delayed, the idle time period within which the machine can finish the working procedure can be searched according to the earliest possible work time of the workpiece to perform plug-in decoding. In the existing job shop scheduling, the earliest working time of greedy decoding processing of a workpiece is the completion time of a preorder process; the invention performs more constraints on the greedy decoding process based on the constraint conditions, for example, the earliest start-up time of the workpiece should be the end time of the AGV load. By means of the constraint conditions, greedy decoding of whale individuals in whale populations can be performed more smoothly and rapidly in combination with the running paths of the AGV trolleys and the corresponding transportation running time, and the decoding results can meet the process constraint requirements of actual processing operation.

In the step 6) of carrying out whale optimization algorithm iterative updating on the whale population, the whale population can be processed according to the existing whale optimization algorithm iterative updating operation. However, as a preferred scheme of the method, on the basis of whale optimization algorithm iterative updating, levy flight operator updating can be introduced by adding one's complement, so that the global search capability of the algorithm is enhanced, and the capability of the algorithm for jumping out of local optimum is also enhanced. Because the convergence factor a of the whale optimization algorithm is gradually reduced in the later iteration stage, the algorithm is in risk of being trapped in local optimization. Levy flight is a random search path that follows a Levy distribution. Therefore, the Levy flight strategy is introduced into algorithm iteration, so that the capacity of jumping out of local optimum at the later stage can be enhanced; meanwhile, the global search capability of the algorithm can be enhanced by introducing a Levy flight strategy in the early stage of the algorithm.

The specific mode of introducing Levy flight operators to carry out whale optimization algorithm iterative updating on whale populations is as follows:

updating a calculated coefficient vector

The value of (c):

convergence factor a =2- (2 t)/t _max ；/>

Wherein t is the current iteration number, t _max Is the maximum number of iterations, r ₁ 、r ₂ Are all [0,1]A random number taken between; in [0,1]Randomly taking values to generate a probability parameter p;

when coefficient vector

Is greater than or equal to>

the first method is as follows:

when coefficient vector

Is greater than or equal to>

And when p is more than or equal to 0.5, updating the individual positions of the whales according to the following mode II: />

The second method comprises the following steps:

when coefficient vector

Is greater than or equal to>

the third method comprises the following steps:

wherein the content of the first and second substances,

Indicating the individual position, in conjunction with the sperm of a present whale>

Indicating a target prey location; b is a defined logarithmic spiral shape constant; l is [ -1,1]A random number taken between; rand is subject to a range of [0,1]A uniform distribution function of; when/is>

Then, [ rand-1/2 ]]Taking-1; when/is>

Then, [ rand-1/2 ]]Taking 0; when/is>

Then, [ rand-1/2 ]]Taking 1; />

Levy(s)～|s| ^-1-β ,0＜β≤2；

wherein u, v and beta are Levy flight parameters; beta is more than 0 and less than or equal to 2, and the beta is 1.5 generally; u and v are normal distribution random numbers; u obeys normal distribution

σ _u As shown in equation (9), Γ is the standard gamma function; v obey a normal distribution>

σ _v I.e., v obeys a normal distribution v to N (0,1).

Threshold operations are also one of the common practices today to prevent algorithms from falling into local optimality. The specific method comprises the following steps: setting a global optimal whale individual iteration number gic with the initial value equal to zero when the algorithm starts, recording an iteration number value with the global optimal whale individual unchanged in the algorithm iteration process, and replacing 50% of inferior whale individuals in the whale population with generated new whale individuals when gic reaches a preset algebraic threshold lim. In the invention, the quality of individual whales is judged according to the maximum completion time, namely, the individual whales are judged to be inferior when the maximum completion time is longer, and the individual whales are judged to be superior when the maximum completion time is shorter. Here, the value of the algebraic threshold has a large influence on the algorithm; if the threshold is set to be too small, the whale individual is not completely converged, and the complexity of the algorithm is increased rapidly; if the threshold is set too large, the convergence speed of the algorithm is slow. The value of the algebraic threshold is generally set to be 10% -15% of the maximum iteration number, and is a better value selection.

For the local search processing, in the development stage, the population individuals update the positions of the population individuals by taking the current optimal individuals as the reference, and the local search is performed on the better individuals to improve the quality of the better individuals, so that the algorithm solving precision and the convergence speed can be greatly improved. Because AGV devices are relatively expensive, the number of AGVs in a plant is often low. For the problem of co-scheduling of a machine and an AGV, a phenomenon that a workpiece waits for the AGV to transport often occurs, and reasonable distribution of the AGV is very important for reducing the maximum completion time. The AGV that completes the transfer task at the latest is defined as the restraint AGV. In the invention, the local search processing scheme as the improved optimization is divided into two parts of processing links: variable neighborhood search based on process and machine, and neighborhood search based on constrained AGVs.

The variable neighborhood searching algorithm enables a searching space to be wider and deeper by progressively checking different neighborhoods, and has stronger local searching capability. Aiming at the condition that workpieces wait for AGV transportation frequently occurring in the problem of co-scheduling of a machine and an AGV, in order to better optimize solution, the method and the system for scheduling whales in the whale population at present _x Individual superior whales are searched locally, n _x A number parameter for a preset local search, and n _x The value of (2) is an integer greater than 1 and less than or equal to 5, and the probability of optimizing the current solution can be increased relative to the variable neighborhood search of only 1 optimal individual; the local search number parameter n can be taken _x Has a value of 3.

Designing a variable neighborhood searching algorithm, firstly, a neighborhood structure is designed. The invention preferably adopts the following three neighborhood structures:

Based on the three neighborhood structures, the local search processing flow for each whale individual needs to go through two links of variable neighborhood search based on processes and machines and neighborhood search based on restricted AGV.

The variable neighborhood searching based on the process and the machine comprises the following steps:

step a1: firstly, X is _a Updating the whale individuals subjected to local search currently; let the current iteration number n _a =1, let maximum number of iterations n _a,max =5, let p _a =1, let p _a,max ＝3；

Step a2: judging whether a circulation termination condition n is reached _a ≥n _a,max (ii) a If it reaches, output X _a Corresponding individual whale; otherwise, turning to the step a3;

step a3: at X _a Randomly selecting a neighborhood structure on the basis of the corresponding whale individual to obtain a disturbance individual X' _a ；

Step a4: in perturbed individual X' _a The variable neighborhood searching is carried out on the basis, and the specific steps are as follows:

a4.2 X 'in a perturbed individual' _a Based on the selection number and the current p _a Get the new individual X ″' with the neighborhood structure corresponding to the value _a If f (X ″) _a )＜f(X′ _a ) Then X 'is updated' _a ←X″ _a ，p _a Axle 300, 1; if f (X ″) _a )＝f(X′ _a ') then X ' is updated with a probability of 0.5 ' _a ←X″ _a ，p _a Axle 300, 1; otherwise X' _a Not updated, p _a ←p _a +1; then returning to a 4.1);

step a5: updating X _a ←X′ _a ，n _a ←n _a +1, go to step a2.

Wherein, f (X' _a )、f(X″ _a ) Respectively represent individual X' _a 、X″ _a The maximum completion time of the respective machine processing.

Then, based on the neighborhood search of the constraint AGV, an individual output by variable neighborhood search based on processes and machines is used as an initial individual value, and the method comprises the following steps:

step b1: x is to be _b Updating X output for the variable neighborhood search based on process and machine _a A corresponding individual whale; order the AGV task number n of the current operation _b Is 1, let n _b,max The total task number of the AGV currently operating;

step b3: distributing the AGV task operated currently to the AGV trolley with the minimum current transportation running time to obtain a new individual X' _b (ii) a If f (X' _b )＜f(X _b ) Update X _b ←X′ _b (ii) a If f (X' _b )＝f(X _b ) Then X is updated with a probability of 0.5 _b ←X′ _b (ii) a Otherwise, X is not updated _b (ii) a Then n is updated _b ←n _b +1, go to step b2.

Wherein f (X' _b )、f(X″ _b ) Respectively represent individual X' _b 、X″ _b The respective AGVs transport the maximum completion time.

For the AGV path planning problem, in the process of decoding an individual by an algorithm, after a process determines a processing machine, the actual completion time of the process needs to be determined according to the earliest possible work time of a workpiece and the processing time of the machine. In the traditional workshop scheduling scheme, the earliest starting time of a process is the completion time of the previous process, but the integrated scheduling problem of the AGV and the machine, which is provided by the invention, needs to consider the transfer time of a workpiece and the path conflict possibly encountered in the path, so that a conflict-free path needs to be planned for the AGV in order to determine the earliest starting time of the process. In order to implement collision-free path planning for AGVs, the existing collision-free path planning methods such as Dijkstra algorithm based on time windows can be adopted to implement ordered path planning for AGVs and reduce or prevent the problem of path collision, for example, the technical scheme of AGV path planning patents such as CN106251016A, CN107167154A, CN107179773A in our country.

Flexible example experiment

In order to verify the feasibility and the effectiveness of the discrete whale optimization algorithm provided by the invention, an example in a reference of 'Lyu X, song Y, he C, et al, application to Integrated Scheduling schemes configuring Optime Number of automatic Guided Vehicles and communicating-Free Routing in Flexible Manufacturing Systems [ J ]. IEEE Access,2019, 74909-74924' is selected for testing, the example comprises 4 workpieces and 6 machines, and the Number of AGVs is increased from 1 to 6. The PGA operating parameters in this reference are: the population size is set to 80, the maximum iteration number is set to 80, the crossing rate is set to 0.9, the variation rate is set to 0.1, and the algorithm is independently run for 10 times to obtain the best result. For better comparison, the population size, maximum number of iterations, number of independent runs, and the reference were set to the same values for the method of the present invention (labeled IWOA in this example) and the original whale optimization algorithm of the prior art (labeled WOA in this example). The whale individual boundary of WOA is set to 1, the whale individual boundary of iwoa is set to 3, and the optimal whale individual retention algebra threshold is set to 10.

In Table 1, v represents the AGV number, C _max Represents the maximum completion time in minutes and the CPU represents the run time in seconds.

Table 1 results of compliance calculations

As can be seen from table 1, the maximum completion time is greater for the IWOA results than for the PGA, and the maximum completion time is better than for the PGA with the result that the operation time is slightly worse than the PGA. The overall operational results for the maximum completion time of IWOA are better than WOA.

Fig. 6 is a gantt chart when the number of AGVs is 3, and ET represents an empty travel of the AGVs. It can be seen that all the process start time is after the AGV load end time, the workpiece preorder process completion time, and the machine preorder process completion time. And the constraint condition is satisfied.

Fig. 7 is a time window diagram of each road section when the number of AGVs is 3, and AGVs are distinguished by different colors, so that it can be seen that the same AGV does not appear on the ordinate at any time, which illustrates that an AGV allocation scheme is feasible; in any road section, the overlapping of different AGV time windows can not occur, and the AGV can not collide with each other. The algorithm of the present invention can plan a collision-free path for an AGV.

FIG. 8 is a graph of the convergence of WOA and IWOA for an AGV number of 3. The maximum completion time of the WOA in the middle and later stages of iteration has a larger reduction range, so that the local development capability of the WOA in the later stages can be seen, but the maximum completion time is not changed in the later stages, and the characteristic that the WOA is easy to fall into the local optimal solution can be seen; compared with WOA, the quality of the initial population of the IWOA method is higher, and the effectiveness of the method is verified; the IWOA method of the invention starts to converge to 89min in 25 generations, which shows that the algorithm has better convergence speed and convergence precision.

Temporal complexity analysis

The time complexity is an important performance of the algorithm, and can reflect the operation efficiency of the algorithm. Assuming that the population scale of the original whale optimization algorithm is N and the dimensionality of individual whales is N, the time complexity T of the initialization stage ₁ = O (n). Assuming that after iteration, the calculation time of the maximum completion time of an individual is f (n), and the time for processing the dimensionality out-of-range of the individual of the whale is t ₁ The two parts of time complexity T ₂ ＝O(N(f(n)+t ₁ N)), assuming that the number of the three update modes is m ₁ 、m ₂ 、m ₃ Each dimension has an update time t ₂ 、t ₃ 、t ₄ Then the time complexity T in the iterative process ₃ ＝O(m ₁ (n·t ₂ )+m ₂ (n·t ₃ )+m ₃ ·(n·t ₄ )). The time complexity of the original whale optimization algorithm in the prior art is the sum of the three parts, and the time complexity is specifically shown in the following formula:

T _WOA ＝2O(n)+O(n+f(n))＝O(n+f(n))

in the algorithm provided by the invention, the number of machines is assumed to be m, and the time for calculating the load of each machine is t ₅ Time complexity T of GLR selection strategy ₁ ＝O(0.5N·n+0.5N·n·(m·t ₅ ) O (n), assuming the computation time of chaotic mapping and opponent learning is T6 and T7, the process time complexity T ₂ ＝O(N·n·t ₆ +N·t ₇ ). After iteration is carried out, the calculation time of Levy step length is set as t ₈ Location update complexity T ₃ ＝O(m ₁ (n(t ₂ +t ₈ ))+m ₂ (n·t ₃ )+m ₃ (n(t ₄ +t ₈ ))). The maximum completion time and the time to process the boundary crossing are the same as the original whale optimization algorithm. Time complexity T of threshold restart operation ₄ O (0.5N · N) = O (N). Assume that the time to construct a neighborhood is t ₉ Calculating the AGV running time as t ₁₀ Time complexity of local search is T ₅ ＝O(n _max ·q _max ·f(n)·(t ₉ +t ₁₀ )). Thus, the time complexity of the method of the invention is as follows:

T _IWOA ＝3O(n)+O(n+f(n))+O(f(n))＝O(n+f(n))

compared with the existing original whale optimization algorithm, the algorithm has the advantages that the complexity of the operation processing time of the algorithm is the same, and the operation time under the same operation environment is in the same order of magnitude. Therefore, it is also demonstrated that, while the method of the present invention realizes effective solution of the integrated scheduling problem of AGVs and machines, compared with the existing original whale optimization algorithm, the method of the present invention not only can enhance the global search capability of the algorithm and the capability of jumping out of local optimum, and optimize and improve the scheduling efficiency of the integrated operation flow of AGVs and machines, but also does not sacrifice the complexity of operation, still maintains good solution operation efficiency, and has good convergence speed and convergence accuracy.

Finally, it is noted that the above-mentioned embodiments illustrate rather than limit the invention, and that, while the invention has been described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. An AGV and machine integrated scheduling method based on a discrete whale optimization algorithm is characterized by comprising the following steps:

carrying out whale individual coding treatment on the whale optimization algorithm according to the scheduling task parameters to generate a whale population with a preset population scale, wherein each whale individual corresponds to an AGV and machine integrated operation process scheduling scheme; on the basis of randomly generating whale individuals with the population scale of 50%, generating an initialization population by using Gauss mapping and standard opposite learning and respectively generating strategies with the population scale proportion of 25% to obtain whale populations with the preset population scale;

the gaussian mapping expression is:

the standard opponent learning expression is:

wherein the content of the first and second substances,

representing the encoded value of the ith dimension in the individual sperm of a whale after Gaussian mapping]Represents integrate, or de-integrate, based on a predetermined criterion>

Representing whales after opponent learningCoded value for the i-th dimension in an individual fish>

And &>

Coded values ^ for the ith dimension in randomly generated individual whales>

Upper and lower limit values of the value;

according to the AGV and machine integrated operation flow scheduling scheme corresponding to the optimal whale individual, performing operation flow scheduling of the AGV trolley and the machine;

the whale individual comprises a plurality of genes, each gene corresponds to one operation process in the AGV and machine integrated operation flow scheduling scheme, and therefore a whale individual is formed by a set of genes corresponding to the operation processes in the AGV and machine integrated operation flow scheduling scheme; the whale individual codes of all genes in the whale individuals comprise procedure code segments, machine code segments and AGV code segments, wherein the procedure code segments are used for indicating workpiece numbers corresponding to the operation procedures, the machine code segments are used for indicating machine numbers corresponding to the workpiece numbers indicated in the procedure code segments in the operation procedures, and the AGV code segments are used for indicating AGV numbers corresponding to the workpiece numbers indicated in the procedure code segments in the operation procedures; and in a whale individual, the repeated sequence of the genes with the same process code segment is used for indicating the processing process of the workpiece corresponding to the workpiece number in the corresponding process code segment;

each code segment of the gene in the whale individual is continuous coding with a value interval of [ -delta, delta ], and discrete coding is obtained after discrete coding conversion so as to indicate a workpiece number, a machine number and an AGV number corresponding to the gene;

the discrete transcoding method for individual whale is as follows:

performing discrete code conversion by adopting a maximum position value rule aiming at the process code segment of each gene in the whale individual; the discrete code conversion processing process comprises the following steps: firstly, setting a fixed ID corresponding to each gene in whale individuals in an increasing order from 1, then sequencing and rearranging each gene in the whale individuals according to the sequence of the values of the process code segments from large to small, disorder the fixed ID sequencing of each gene, forming mapping ID sequencing, and obtaining a new gene sequencing sequence; then, according to the sequence of the values of the mapping IDs from small to large, inserting the procedures of the workpiece 1 into the gene positions in the corresponding sequence in sequence, and then inserting the procedures of the workpiece 2; by analogy, until all the procedures of each workpiece are inserted, the inserted value is the workpiece serial number value corresponding to the procedure, and therefore the discrete coding sequence of the procedure code segment in the whale individual is obtained;

m _equ (i) Representing continuously encoded machine code segment values, z _equ (i) Indicates the number of processing machines, u, selectable for the respective process _equ (i) Representing the resulting discretely encoded machine code segment values; m is a unit of _AGV (i) Representing successively encoded AGV segment values, z _AGV (i) Indicates the number of AGV carts selectable for the corresponding process, u _AGV (i) Representing the obtained discrete coded AGV code segment value; delta is the value range of continuous coding [ -delta, delta]The upper limit end value of (2); run [ 2 ]]Rounding operator;

in the iterative process of whale population, after a whale individual is subjected to position updating, the whale individual is subjected to border crossing adjustment, and the adjustment rule is shown as the following formula:

r ₁ represents [0,1]The value of the whale individual X is a random number, and X (i) represents the coded value of the ith dimension of the whale individual X;

the specific process of the method comprises the following steps:

2) Carrying out whale individual coding treatment of a whale optimization algorithm according to the scheduling task parameters to generate whale populations with the NIND whale individual scales;

3) Judging whether the iteration loop times reach the maximum iteration times, and if so, skipping to execute the step 9); otherwise, planning running paths of all AGV dollies in the integrated operation flow scheduling scheme of the AGVs and the machines corresponding to all whale individuals in the whale population, determining the transportation running time of each AGV dollies in the planned path, carrying out greedy decoding on each whale individual in the whale population by combining the running paths of the AGV dollies and the corresponding transportation running time, respectively calculating the maximum completion time of the integrated operation flow scheduling scheme of the AGVs and the machines corresponding to each whale individual in the whale population, recording the whale individual with the shortest current global maximum completion time as a global optimal whale individual, and executing the step 4); in step 3), the constraint condition for greedy decoding of individual whales in the whale population comprises:

constraint (1): any one procedure can be processed only by one machine;

constraint (11): any position node can only accommodate the next AGV at the same time;

6) Carrying out whale optimization algorithm iterative updating on the whale population; in the step 6), a specific mode of carrying out whale optimization algorithm iterative updating on the whale population is as follows:

updating a vector of computed coefficients

The value of (c):

convergence factor a =2- (2 t)/t _max ；

when coefficient vector

Is greater than or equal to>

the first method is as follows:

when coefficient vector

Is greater than or equal to>

And when p is more than or equal to 0.5, updating the individual positions of the whales according to the following mode II:

the second method comprises the following steps:

when coefficient vector

Is greater than or equal to>

the third method comprises the following steps:

wherein the content of the first and second substances,

Indicating a target prey location; b is a defined logarithmic spiral shape constant; l is [ -1,1]A random number taken between; rand is subject to a range of [0,1]A uniform distribution function of; when +>

Then, [ rand-1/2 ]]Taking-1; when +>

Then, [ rand-1/2 ]]Taking 0; when/is>

Then, [ rand-1/2 ]]Taking 1; />

Levy(s)～|s| ^-1-β ,0＜β≤2；

/>

σ _v ＝1；

u, v and beta are Levy flight parameters; u obeys normal distribution

v obey a normal distribution>

Gamma is a standard gamma function;

7) For n in the current whale population _x Carrying out local search on individual superior whales; n is _x A preset local search quantity parameter; in step 7), the neighborhood structure adopted by the local search comprises:

neighborhood structure 3: randomly selecting one of the process code segments of each gene in the whale individual, wherein the number of machinable machines in the process corresponding to the process code segment is more than 1, and randomly selecting one machining machine for replacing the process corresponding to the process code segment from the machinable machines in the process corresponding to the process code segment;

the specific way of carrying out local search on each whale individual is as follows:

a4.2 X 'in a perturbed individual' _a Based on the selection number and the current p _a Get the new individual X ″' with the neighborhood structure corresponding to the value _a If f (X ″) _a )＜f(X′ _a ) Then X 'is updated' _a ←X″ _a ，p _a Axle 300, 1; if f (X ″) _a )＝f(X′ _a ) Then X 'is updated with a probability of 0.5' _a ←X″ _a ，p _a Axle 300, 1; otherwise X' _a Not updated, p _a ←p _a +1; then return to a 4.1);

step a5: updating X _a ←X′ _a ，n _a ←n _a +1, turning to step a2;

step b1: x is to be _b Updating X output for the variable neighborhood search based on process and machine _a A corresponding individual whale; let the currently operating AGV task number n _b Is 1, let n _b,max The total task number of the AGV currently operating;

and b2: judging whether a termination condition n is satisfied _b ≥n _b,max If satisfied, output X _b The corresponding whale individual is used as a preferred object of the whale population in the step 8); otherwise, turning to the step b3;

step b3: distributing the AGV task operated currently to the AGV trolley with the minimum current transportation running time to obtain a new individual X' _b (ii) a If f (X' _b )＜f(X _b ) Update X _b ←X′ _b (ii) a If f (X' _b )＝f(X _b ) Then X is updated with a probability of 0.5 _b ←X′ _b (ii) a Otherwise, X is not updated _b (ii) a Then n is updated _b ←n _b +1, go to step b2;

wherein, f (X' _b )、f(X″ _b ) Respectively represent individual X' _b 、X″ _b Respectively corresponding AGV transportation maximum completion time;

8) Keeping the whale population scale, selecting a better whale individual, performing the next iteration, and returning to the step 3);

2. The discrete whale optimization algorithm-based AGV and machine integrated scheduling method according to claim 1, wherein the scheduling task parameters comprise processing parameters of workpieces, transport operation parameters of AGV carts, and map information of transfer transport areas in a work flow.

3. The AGV and machine integrated scheduling method based on discrete whale optimization algorithm of claim 1, wherein in the step 3), the whale individual with the shortest current global maximum completion time is the whale individual with the shortest maximum completion time in each iteration from the current iteration time.